Merge branch 'jc/blame' (early part)
[git/git.git] / dir.c
CommitLineData
453ec4bd
LT
1/*
2 * This handles recursive filename detection with exclude
3 * files, index knowledge etc..
4 *
5 * Copyright (C) Linus Torvalds, 2005-2006
6 * Junio Hamano, 2005-2006
7 */
453ec4bd
LT
8#include "cache.h"
9#include "dir.h"
09595258 10#include "refs.h"
453ec4bd 11
9fc42d60
LT
12struct path_simplify {
13 int len;
14 const char *path;
15};
16
09595258
LT
17static int read_directory_recursive(struct dir_struct *dir,
18 const char *path, const char *base, int baselen,
19 int check_only, const struct path_simplify *simplify);
20
3c6a370b
LT
21int common_prefix(const char **pathspec)
22{
23 const char *path, *slash, *next;
24 int prefix;
25
26 if (!pathspec)
27 return 0;
28
29 path = *pathspec;
30 slash = strrchr(path, '/');
31 if (!slash)
32 return 0;
33
34 prefix = slash - path + 1;
35 while ((next = *++pathspec) != NULL) {
36 int len = strlen(next);
c7f34c18 37 if (len >= prefix && !memcmp(path, next, prefix))
3c6a370b 38 continue;
c7f34c18 39 len = prefix - 1;
3c6a370b
LT
40 for (;;) {
41 if (!len)
42 return 0;
43 if (next[--len] != '/')
44 continue;
45 if (memcmp(path, next, len+1))
46 continue;
47 prefix = len + 1;
48 break;
49 }
50 }
51 return prefix;
52}
53
e813d50e
JH
54/*
55 * Does 'match' matches the given name?
56 * A match is found if
57 *
58 * (1) the 'match' string is leading directory of 'name', or
59 * (2) the 'match' string is a wildcard and matches 'name', or
60 * (3) the 'match' string is exactly the same as 'name'.
61 *
62 * and the return value tells which case it was.
63 *
64 * It returns 0 when there is no match.
65 */
3c6a370b
LT
66static int match_one(const char *match, const char *name, int namelen)
67{
68 int matchlen;
69
70 /* If the match was just the prefix, we matched */
71 matchlen = strlen(match);
72 if (!matchlen)
e813d50e 73 return MATCHED_RECURSIVELY;
3c6a370b
LT
74
75 /*
76 * If we don't match the matchstring exactly,
77 * we need to match by fnmatch
78 */
79 if (strncmp(match, name, matchlen))
e813d50e 80 return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0;
3c6a370b 81
e813d50e
JH
82 if (!name[matchlen])
83 return MATCHED_EXACTLY;
84 if (match[matchlen-1] == '/' || name[matchlen] == '/')
85 return MATCHED_RECURSIVELY;
86 return 0;
3c6a370b
LT
87}
88
e813d50e
JH
89/*
90 * Given a name and a list of pathspecs, see if the name matches
91 * any of the pathspecs. The caller is also interested in seeing
92 * all pathspec matches some names it calls this function with
93 * (otherwise the user could have mistyped the unmatched pathspec),
94 * and a mark is left in seen[] array for pathspec element that
95 * actually matched anything.
96 */
3c6a370b
LT
97int match_pathspec(const char **pathspec, const char *name, int namelen, int prefix, char *seen)
98{
99 int retval;
100 const char *match;
101
102 name += prefix;
103 namelen -= prefix;
104
105 for (retval = 0; (match = *pathspec++) != NULL; seen++) {
e813d50e
JH
106 int how;
107 if (retval && *seen == MATCHED_EXACTLY)
3c6a370b
LT
108 continue;
109 match += prefix;
e813d50e
JH
110 how = match_one(match, name, namelen);
111 if (how) {
112 if (retval < how)
113 retval = how;
114 if (*seen < how)
115 *seen = how;
3c6a370b
LT
116 }
117 }
118 return retval;
119}
120
453ec4bd
LT
121void add_exclude(const char *string, const char *base,
122 int baselen, struct exclude_list *which)
123{
124 struct exclude *x = xmalloc(sizeof (*x));
125
126 x->pattern = string;
127 x->base = base;
128 x->baselen = baselen;
129 if (which->nr == which->alloc) {
130 which->alloc = alloc_nr(which->alloc);
83572c1a
JF
131 which->excludes = xrealloc(which->excludes,
132 which->alloc * sizeof(x));
453ec4bd
LT
133 }
134 which->excludes[which->nr++] = x;
135}
136
137static int add_excludes_from_file_1(const char *fname,
138 const char *base,
139 int baselen,
140 struct exclude_list *which)
141{
c470701a 142 struct stat st;
453ec4bd 143 int fd, i;
dc49cd76 144 size_t size;
453ec4bd
LT
145 char *buf, *entry;
146
147 fd = open(fname, O_RDONLY);
c470701a 148 if (fd < 0 || fstat(fd, &st) < 0)
453ec4bd 149 goto err;
dc49cd76 150 size = xsize_t(st.st_size);
453ec4bd
LT
151 if (size == 0) {
152 close(fd);
153 return 0;
154 }
155 buf = xmalloc(size+1);
93d26e4c 156 if (read_in_full(fd, buf, size) != size)
453ec4bd
LT
157 goto err;
158 close(fd);
159
160 buf[size++] = '\n';
161 entry = buf;
162 for (i = 0; i < size; i++) {
163 if (buf[i] == '\n') {
164 if (entry != buf + i && entry[0] != '#') {
165 buf[i - (i && buf[i-1] == '\r')] = 0;
166 add_exclude(entry, base, baselen, which);
167 }
168 entry = buf + i + 1;
169 }
170 }
171 return 0;
172
173 err:
174 if (0 <= fd)
175 close(fd);
176 return -1;
177}
178
179void add_excludes_from_file(struct dir_struct *dir, const char *fname)
180{
181 if (add_excludes_from_file_1(fname, "", 0,
182 &dir->exclude_list[EXC_FILE]) < 0)
183 die("cannot use %s as an exclude file", fname);
184}
185
f8a9d428 186int push_exclude_per_directory(struct dir_struct *dir, const char *base, int baselen)
453ec4bd
LT
187{
188 char exclude_file[PATH_MAX];
189 struct exclude_list *el = &dir->exclude_list[EXC_DIRS];
190 int current_nr = el->nr;
191
192 if (dir->exclude_per_dir) {
193 memcpy(exclude_file, base, baselen);
194 strcpy(exclude_file + baselen, dir->exclude_per_dir);
195 add_excludes_from_file_1(exclude_file, base, baselen, el);
196 }
197 return current_nr;
198}
199
f8a9d428 200void pop_exclude_per_directory(struct dir_struct *dir, int stk)
453ec4bd
LT
201{
202 struct exclude_list *el = &dir->exclude_list[EXC_DIRS];
203
204 while (stk < el->nr)
205 free(el->excludes[--el->nr]);
206}
207
208/* Scan the list and let the last match determines the fate.
209 * Return 1 for exclude, 0 for include and -1 for undecided.
210 */
211static int excluded_1(const char *pathname,
212 int pathlen,
213 struct exclude_list *el)
214{
215 int i;
216
217 if (el->nr) {
218 for (i = el->nr - 1; 0 <= i; i--) {
219 struct exclude *x = el->excludes[i];
220 const char *exclude = x->pattern;
221 int to_exclude = 1;
222
223 if (*exclude == '!') {
224 to_exclude = 0;
225 exclude++;
226 }
227
228 if (!strchr(exclude, '/')) {
229 /* match basename */
230 const char *basename = strrchr(pathname, '/');
231 basename = (basename) ? basename+1 : pathname;
232 if (fnmatch(exclude, basename, 0) == 0)
233 return to_exclude;
234 }
235 else {
236 /* match with FNM_PATHNAME:
237 * exclude has base (baselen long) implicitly
238 * in front of it.
239 */
240 int baselen = x->baselen;
241 if (*exclude == '/')
242 exclude++;
243
244 if (pathlen < baselen ||
245 (baselen && pathname[baselen-1] != '/') ||
246 strncmp(pathname, x->base, baselen))
247 continue;
248
249 if (fnmatch(exclude, pathname+baselen,
250 FNM_PATHNAME) == 0)
251 return to_exclude;
252 }
253 }
254 }
255 return -1; /* undecided */
256}
257
258int excluded(struct dir_struct *dir, const char *pathname)
259{
260 int pathlen = strlen(pathname);
261 int st;
262
263 for (st = EXC_CMDL; st <= EXC_FILE; st++) {
264 switch (excluded_1(pathname, pathlen, &dir->exclude_list[st])) {
265 case 0:
266 return 0;
267 case 1:
268 return 1;
269 }
270 }
271 return 0;
272}
273
6815e569 274static struct dir_entry *dir_entry_new(const char *pathname, int len) {
453ec4bd
LT
275 struct dir_entry *ent;
276
453ec4bd 277 ent = xmalloc(sizeof(*ent) + len + 1);
4d06f8ac 278 ent->ignored = ent->ignored_dir = 0;
453ec4bd
LT
279 ent->len = len;
280 memcpy(ent->name, pathname, len);
281 ent->name[len] = 0;
4d06f8ac 282 return ent;
453ec4bd
LT
283}
284
6815e569
JK
285struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
286{
287 if (cache_name_pos(pathname, len) >= 0)
288 return NULL;
289
290 ALLOC_GROW(dir->entries, dir->nr, dir->alloc);
291 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
292}
293
09595258
LT
294enum exist_status {
295 index_nonexistent = 0,
296 index_directory,
297 index_gitdir,
298};
299
300/*
301 * The index sorts alphabetically by entry name, which
302 * means that a gitlink sorts as '\0' at the end, while
303 * a directory (which is defined not as an entry, but as
304 * the files it contains) will sort with the '/' at the
305 * end.
306 */
307static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd
LT
308{
309 int pos = cache_name_pos(dirname, len);
09595258
LT
310 if (pos < 0)
311 pos = -pos-1;
312 while (pos < active_nr) {
313 struct cache_entry *ce = active_cache[pos++];
314 unsigned char endchar;
315
316 if (strncmp(ce->name, dirname, len))
317 break;
318 endchar = ce->name[len];
319 if (endchar > '/')
320 break;
321 if (endchar == '/')
322 return index_directory;
302b9282 323 if (!endchar && S_ISGITLINK(ntohl(ce->ce_mode)))
09595258
LT
324 return index_gitdir;
325 }
326 return index_nonexistent;
327}
328
329/*
330 * When we find a directory when traversing the filesystem, we
331 * have three distinct cases:
332 *
333 * - ignore it
334 * - see it as a directory
335 * - recurse into it
336 *
337 * and which one we choose depends on a combination of existing
338 * git index contents and the flags passed into the directory
339 * traversal routine.
340 *
341 * Case 1: If we *already* have entries in the index under that
342 * directory name, we always recurse into the directory to see
343 * all the files.
344 *
345 * Case 2: If we *already* have that directory name as a gitlink,
346 * we always continue to see it as a gitlink, regardless of whether
347 * there is an actual git directory there or not (it might not
348 * be checked out as a subproject!)
349 *
350 * Case 3: if we didn't have it in the index previously, we
351 * have a few sub-cases:
352 *
353 * (a) if "show_other_directories" is true, we show it as
354 * just a directory, unless "hide_empty_directories" is
355 * also true and the directory is empty, in which case
356 * we just ignore it entirely.
357 * (b) if it looks like a git directory, and we don't have
302b9282 358 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
359 * as a directory.
360 * (c) otherwise, we recurse into it.
361 */
362enum directory_treatment {
363 show_directory,
364 ignore_directory,
365 recurse_into_directory,
366};
367
368static enum directory_treatment treat_directory(struct dir_struct *dir,
369 const char *dirname, int len,
370 const struct path_simplify *simplify)
371{
372 /* The "len-1" is to strip the final '/' */
373 switch (directory_exists_in_index(dirname, len-1)) {
374 case index_directory:
375 return recurse_into_directory;
376
377 case index_gitdir:
ab22aed3
LT
378 if (dir->show_other_directories)
379 return ignore_directory;
09595258
LT
380 return show_directory;
381
382 case index_nonexistent:
383 if (dir->show_other_directories)
384 break;
302b9282 385 if (!dir->no_gitlinks) {
09595258
LT
386 unsigned char sha1[20];
387 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
388 return show_directory;
389 }
390 return recurse_into_directory;
391 }
392
393 /* This is the "show_other_directories" case */
394 if (!dir->hide_empty_directories)
395 return show_directory;
396 if (!read_directory_recursive(dir, dirname, dirname, len, 1, simplify))
397 return ignore_directory;
398 return show_directory;
453ec4bd
LT
399}
400
9fc42d60
LT
401/*
402 * This is an inexact early pruning of any recursive directory
403 * reading - if the path cannot possibly be in the pathspec,
404 * return true, and we'll skip it early.
405 */
406static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
407{
408 if (simplify) {
409 for (;;) {
410 const char *match = simplify->path;
411 int len = simplify->len;
412
413 if (!match)
414 break;
415 if (len > pathlen)
416 len = pathlen;
417 if (!memcmp(path, match, len))
418 return 0;
419 simplify++;
420 }
421 return 1;
422 }
423 return 0;
424}
425
453ec4bd
LT
426/*
427 * Read a directory tree. We currently ignore anything but
428 * directories, regular files and symlinks. That's because git
429 * doesn't handle them at all yet. Maybe that will change some
430 * day.
431 *
432 * Also, we ignore the name ".git" (even if it is not a directory).
433 * That likely will not change.
434 */
9fc42d60 435static int read_directory_recursive(struct dir_struct *dir, const char *path, const char *base, int baselen, int check_only, const struct path_simplify *simplify)
453ec4bd
LT
436{
437 DIR *fdir = opendir(path);
438 int contents = 0;
439
440 if (fdir) {
441 int exclude_stk;
442 struct dirent *de;
095c424d 443 char fullname[PATH_MAX + 1];
453ec4bd
LT
444 memcpy(fullname, base, baselen);
445
446 exclude_stk = push_exclude_per_directory(dir, base, baselen);
447
448 while ((de = readdir(fdir)) != NULL) {
449 int len;
b9916256 450 int exclude;
453ec4bd
LT
451
452 if ((de->d_name[0] == '.') &&
453 (de->d_name[1] == 0 ||
454 !strcmp(de->d_name + 1, ".") ||
455 !strcmp(de->d_name + 1, "git")))
456 continue;
457 len = strlen(de->d_name);
5d5cea67
LT
458 /* Ignore overly long pathnames! */
459 if (len + baselen + 8 > sizeof(fullname))
460 continue;
453ec4bd 461 memcpy(fullname + baselen, de->d_name, len+1);
9fc42d60
LT
462 if (simplify_away(fullname, baselen + len, simplify))
463 continue;
b9916256
MS
464
465 exclude = excluded(dir, fullname);
466 if (exclude != dir->show_ignored) {
c889763b
JH
467 if (!dir->show_ignored || DTYPE(de) != DT_DIR) {
468 continue;
469 }
470 }
453ec4bd
LT
471
472 switch (DTYPE(de)) {
473 struct stat st;
453ec4bd
LT
474 default:
475 continue;
476 case DT_UNKNOWN:
477 if (lstat(fullname, &st))
478 continue;
479 if (S_ISREG(st.st_mode) || S_ISLNK(st.st_mode))
480 break;
481 if (!S_ISDIR(st.st_mode))
482 continue;
483 /* fallthrough */
484 case DT_DIR:
485 memcpy(fullname + baselen + len, "/", 2);
486 len++;
09595258
LT
487 switch (treat_directory(dir, fullname, baselen + len, simplify)) {
488 case show_directory:
b9916256
MS
489 if (exclude != dir->show_ignored)
490 continue;
453ec4bd 491 break;
09595258
LT
492 case recurse_into_directory:
493 contents += read_directory_recursive(dir,
494 fullname, fullname, baselen + len, 0, simplify);
495 continue;
496 case ignore_directory:
497 continue;
453ec4bd 498 }
09595258 499 break;
453ec4bd
LT
500 case DT_REG:
501 case DT_LNK:
502 break;
503 }
453ec4bd 504 contents++;
07ccbff8
JS
505 if (check_only)
506 goto exit_early;
507 else
4d06f8ac 508 dir_add_name(dir, fullname, baselen + len);
453ec4bd 509 }
07ccbff8 510exit_early:
453ec4bd
LT
511 closedir(fdir);
512
513 pop_exclude_per_directory(dir, exclude_stk);
514 }
515
516 return contents;
517}
518
519static int cmp_name(const void *p1, const void *p2)
520{
521 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
522 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
523
524 return cache_name_compare(e1->name, e1->len,
525 e2->name, e2->len);
526}
527
9fc42d60
LT
528/*
529 * Return the length of the "simple" part of a path match limiter.
530 */
531static int simple_length(const char *match)
453ec4bd 532{
9fc42d60
LT
533 const char special[256] = {
534 [0] = 1, ['?'] = 1,
535 ['\\'] = 1, ['*'] = 1,
536 ['['] = 1
537 };
538 int len = -1;
539
540 for (;;) {
541 unsigned char c = *match++;
542 len++;
543 if (special[c])
544 return len;
545 }
546}
547
548static struct path_simplify *create_simplify(const char **pathspec)
549{
550 int nr, alloc = 0;
551 struct path_simplify *simplify = NULL;
552
553 if (!pathspec)
554 return NULL;
555
556 for (nr = 0 ; ; nr++) {
557 const char *match;
558 if (nr >= alloc) {
559 alloc = alloc_nr(alloc);
560 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
561 }
562 match = *pathspec++;
563 if (!match)
564 break;
565 simplify[nr].path = match;
566 simplify[nr].len = simple_length(match);
567 }
568 simplify[nr].path = NULL;
569 simplify[nr].len = 0;
570 return simplify;
571}
572
573static void free_simplify(struct path_simplify *simplify)
574{
575 if (simplify)
576 free(simplify);
577}
578
579int read_directory(struct dir_struct *dir, const char *path, const char *base, int baselen, const char **pathspec)
580{
581 struct path_simplify *simplify = create_simplify(pathspec);
582
b4189aa8
LT
583 /*
584 * Make sure to do the per-directory exclude for all the
585 * directories leading up to our base.
586 */
587 if (baselen) {
588 if (dir->exclude_per_dir) {
589 char *p, *pp = xmalloc(baselen+1);
590 memcpy(pp, base, baselen+1);
591 p = pp;
592 while (1) {
593 char save = *p;
594 *p = 0;
595 push_exclude_per_directory(dir, pp, p-pp);
596 *p++ = save;
597 if (!save)
598 break;
599 p = strchr(p, '/');
600 if (p)
601 p++;
602 else
603 p = pp + baselen;
604 }
605 free(pp);
606 }
607 }
608
9fc42d60
LT
609 read_directory_recursive(dir, path, base, baselen, 0, simplify);
610 free_simplify(simplify);
453ec4bd
LT
611 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
612 return dir->nr;
613}
c91f0d92
JK
614
615int
616file_exists(const char *f)
617{
618 struct stat sb;
619 return stat(f, &sb) == 0;
620}